Hi,

I am trying to run a task-related correlation analysis following "Simple Correlation Analysis" steps.
In the first step, when I remove effects of no interest (with 3dSynthesize selecting baseline) from my original dataset, I was wondering what happens to the residuals in the dataset? I assume the original dataset = fitts + errts and baseline term is included in the fitts, but the remaining subset after subtraction would still leave me with error terms.
When I remove the effects of no interest, I should subtract it from the fitts dataset instead of the original dataset that is used for the first 3dDeconvolve step (to get cbucket)?

Any help would be greatly appreciated!

Thanks!
SJ